randomly seeing i/o timeouts
delivey
PROOP

a month ago

hey, recently started seeing i/o timeouts after further investigations seems like railway is not sending final ACK back to server for TCP connection initialization. It was fine for months and randomly started failing 15 minutes ago, what could cause this?

Locally it works fine still.

Solved

32 Replies

a month ago

Can you provide more information here?


delivey
PROOP

a month ago

yep, what do you need?


delivey
PROOP

a month ago

seems like it resolved itself, can ignore for now


delivey
PROOP

a month ago

still looking ito it


delivey
PROOP

a month ago

newest issue shows same


erikskrt
PRO

a month ago

@Brody we just had this issue and project id was dfca2de8-200c-4fbb-b74c-74a8a8ee1296

if you guys want to retroactively look into it later, ik you're having a lot of other service issues right now


a month ago

is this resolved for you?


delivey
PROOP

a month ago

yeah its resolved for us


delivey
PROOP

a month ago

I work with him


a month ago

was this a problem from railway->external or external->railway?


delivey
PROOP

a month ago

railway->external


delivey
PROOP

a month ago

based on my investigation it wasnt sending the final ack for the 3-way tcp connection


delivey
PROOP

a month ago

i was able to send requests, etc fine to external from my local


a month ago

which service was it?


delivey
PROOP

a month ago

it wasnt a dns issue because from logs i saw it resolved the ip fine and even tried connecting directly to ip on our staging service


delivey
PROOP

a month ago

can i dm you


delivey
PROOP

a month ago

dont wanna share name here


delivey
PROOP

a month ago

or can i copy service id somehow


a month ago

if you could link it to me that'd be ideal



a month ago

Do you know which service/domain/IP you were trying to reach?


delivey
PROOP

a month ago

yep


delivey
PROOP

a month ago

is there a more private way for us to share


delivey
PROOP

a month ago

internal service domain


delivey
PROOP

a month ago

can also check logs its obvious there


delivey
PROOP

a month ago

17:45 utc issue resolved


delivey
PROOP

a month ago

started around 16:45 utc


delivey
PROOP

a month ago

1468308904290292000


delivey
PROOP

a month ago

@Phineas is there an update on this? we want to figure out if its a railway issue, multiple users saying same points to it, but no official response from Railway


delivey
PROOP

a month ago

kind of a big deal to have external requests failing for an hr


a month ago

Hello, I've done some digging and all of the reports (including yours) were an issue between us and the OVH network specifically. From our perspective, it looks like they had a small network outage at that time.

1468641837542805800


a month ago

also related: https://discord.com/channels/713503345364697088/1468302298626134248 and yeah that external API was using OVH.


Status changed to Solved brody about 1 month ago


Loading...